Kenneth I. Forster - US grants

It has been shown that very briefly presented words can be read and understood, even though they are so brief that we are totally unaware of their presence. The evidence for this is that decisions about whether a sequence of upper-case letters forms a word or not are made more rapidly if it is preceded by a brief presentation of the same sequence, but in lower-case letters (e.g., mystery-MYSTERY). This increase in speed is referred to as priming. It is not simply due to faster recognition of the letters themselves, because there is very little, or no priming when the letter sequences are not words (e.g., bystery-BYSTERY). This suggests that the first sequence somehow primes the neural representation of the second sequence, so that it is more rapidly processed. Recent research using this technique has uncovered an odd fact - priming effects are stronger when the experiment consists of a high proportion of cases where the second sequence is a repetition of the first, rather than being a different sequence (e.g., visitor-MYSTERY). Such a result is not unexpected when both sequences are visible, because the second sequence is highly predictable. However such a result is most unexpected when the first sequence is not perceived. This result might indicate that what we normally take to be conscious processes (e.g., learning to anticipate what the second sequence is likely to be) in fact operate at an unconscious level, and do not require conscious awareness. This would mean that our conscious appreciation of the situation would be irrelevant. With support from the National Science Foundation, Drs. Kenneth Forster and Sachiko Kinoshita are investigating whether this surprising (and disturbing) conclusion is justified, or whether this effect can be explained in simpler ways.

One way to test this conclusion is to train individuals to use a visible cue to predict the correct response to a subsequent stimulus, and then see whether they are able to use that cue when it is so brief that it is no longer visible. Preliminary results indicate that this is not possible. If this can be firmly established, then we know that another explanation is required. One possible explanation is that in some tasks, the difficulty of a given item is a function of the difficulty of the immediately preceding items. If most of the items in the experiment are relatively easy (as they would be when the proportion of repeated sequences is high), then a given item is more likely to be preceded by easy items, and therefore would be responded to more rapidly. Another possibility is that some kind of perceptual learning takes place, in which the visual system develops an efficient procedure for dealing with very rapid sequences of stimuli. This learning would have to be unconscious, but would not necessarily involve processes such as prediction.